Highly publicized reports of infant heparin overdoses occurring in hospitals with state-of-the-art pharmacy programs begs the question of how concentrated heparin vials can wind up in the wrong inventory pocket in an automated dispensing cabinet (ADC). After reviewing several possible scenarios that could lead to ADC stocking errors, we realized that this type of error can (and does) happen every day.
Although dispensing error rates, as a percentage of total dispensed medications, are relatively low, even hospitals using a bar code validation process in the pharmacy experience undetected medication errors every day. It is important to remember that while BCMA processes can offer protection in preventing dispensing errors from reaching patients, it may be years before a majority of small and medium-sized hospitals successfully implement BCMA. To address these concerns, we developed a packaging system using manufacturer’s bulk packaging and custom designed multiple vial packs to improve our dispensing accuracy during the ADC batch-fill process.
ADC Stocking Challenges
Previously, our medication distribution and filling process began with a batch-fill order generated by the ADCs. Pharmacy technicians then retrieved the drugs from the carousel units or from the shelved inventory using bar code verification, counted out the correct number of doses, and placed the order in a zip-lock bag.
Despite having bar code verification at the point of product retrieval—the technician scans the bar code label, storage bin, or shelf—the individual dose is not scanned. Subsequently, when the order is passed to a pharmacist for validation, they will check the label on the outside of the zip-lock bag and quickly review the contents for accuracy, but the fact remains that each dose is not scanned or checked individually during the batch-fill process. Because a 400-bed hospital using a point-of-care distribution system may fill as many as 10,000 doses within a two-hour batch-fill period, changing this system requires significant effort. Even with three pharmacy technicians and three pharmacists to complete the process, each pharmacist must validate the accuracy of over 1,600 doses per hour while attempting to read labels that have critical information printed in a 5- or 6-point font. As such, the failure to carefully check medications prior to dispensing is one of the most common at-risk behaviors in hospital pharmacies.
In 2006, the authors of a study at Brigham and Women’s Hospital concluded that a pharmacy-based, bar code distribution process that does not include scanning every dose of a drug during the dispensing process significantly contributes to the frequency of dispensing errors.1 Even in the face of this evidence, it has proven difficult for pharmacy to meet these scanning expectations due to the limitations in our systems and technology.
Controlled Packaging as a Solution
To avoid these dispensing errors, we looked to controlled packaging to create a safer and more efficient batch-fill process. For this strategy, we defined controlled packaging as packaging that took place under controlled conditions meeting all applicable FDA standards of current Good Manufacturing Practices (cGMPs). This insures that the contents of each package are consistent with the external labeling that identifies the drug name, strength, concentration, vial size, expiration date, and quantity. Examples of controlled packaging include solid oral drugs packaged by a manufacturer in unit dose strips, multiple vial trays sealed in the manufacturer’s original packaging, and smaller multi-unit packs prepared by an FDA-licensed repackager under cGMPs. The goal of our study was to determine whether using controlled packaging for small volume parenteral (SVP) drugs would reduce filling and dispensing errors and if the process would be cost effective.
Testing Packaging Approaches
At SUNY Upstate Medical University Hospital, a 409-bed facility, we dispense approximately four million doses per year from the central inpatient pharmacy department using three automated carousel systems for primary drug storage and ADCs in patient care areas. To begin the process change, we compared our traditional, open-packaging system (zip-lock bags) used to fill medication orders against controlled packaging during the daily morning batch-fill process. Previously, a zip-lock bag would typically contain various quantities of loose vials or a combination of loose vials and manufacturer’s packs. To bridge the gap between larger controlled packages (manufacturer’s packs of 10 or 25 vials) and loose vials, we designed a controlled package—Safe-T-Pack—to contain smaller quantities prepared by an FDA-licensed drug repackaging company under cGMPs. Each Safe-T-Pack would contain five SVP vials. (See Figure 1.)
To perform the comparison, random samples of SVP drug orders packaged in both loose and controlled packaging were reviewed during the technicians’ batch fill and the pharmacists’ validation processes. During this review, a pharmacist checked for the following potential errors:
We excluded SVP drugs that required refrigeration (including vaccines), as well as OR, narcotic, high alert, and look-alike/sound-alike drugs that were ordered in small quantities as the management of these products posed increased logistical considerations.
A batch-fill report was generated every morning before the batch was released in order to identify which SVP drugs were going to be included in that day’s morning batch fill. The number of vials requested by each ADC location was noted so that drug orders retrieved for review after pharmacist validation could be returned to their corresponding delivery bins. Following technician filling and pharmacist verification, random samples of SVP study drug orders were reviewed by the pharmacist, who checked each individual dose for the four error parameters described above.
Table 1 describes the number of SVP drug doses reviewed according to packaging style. Manufacturer’s packs whose original cellophane seal was not broken were considered to be one dose; however, vials that came from the manufacturer in unsealed boxes and vials in Safe-T-Packs were treated as individual vials and thoroughly inspected by the pharmacist reviewer.
The manufacturer’s packs and Safe-T-Packs had a 0% error rate, whereas the zip-lock packaging had a 0.04% error rate. Both wrong dose and expired drugs were found in the zip-lock packaging. All the wrong dose errors were found after the technician fill process, and two expiration date errors were found after the fill process and one expiration date error was found after the pharmacist validation process. (See Table 2.)
Rule of Five
To maintain accurate pharmacy inventory levels with controlled packaging dispensing, changes in the ADC batch-fill process are necessary. At most facilities, ADCs send electronic replenishment orders to the pharmacy when the ADC inventory level for an individual item is at or below the pre-set minimum quantity. At a designated time, the ADC calculates the difference between the actual inventory and the pre-set maximum level to determine the order quantity. As a result of this methodology, pharmacy technicians and pharmacists are often required to fill and dispense single doses of drugs during the ADC replenishment process. This software limitation may be the single largest contributing factor resulting in dispensing and administration errors in hospitals using point-of-care distribution technologies. Simply stated, the more doses of medication each pharmacist has to check, the greater the number of dispensing errors that will occur. If the use of controlled packaging decreased the number of doses that need to be checked for accuracy by even 25%, there would be a statistically significant reduction in dispensing errors and a relative reduction in order validation time.
To address this issue, we created filling strategies for our pharmacy technicians when using controlled packaging. The following rules were implemented for the ADC batch-fill process:
These changes in process made it necessary for our pharmacy technicians to manually adjust the order quantities in our carousels to preserve the accuracy of our pharmacy inventory. This additional step could be eliminated if ADC manufacturers modify their software to allow for the rounding of reorder quantities for select medications. If the batch fill, critical-low and stock-out order quantities automatically rounded to multiples of five, the filling process would be streamlined and reduce the amount of time required for technicians to complete the batch.
When comparing the costs associated with controlled packaging versus those associated with open packaging, it was determined that controlled packaging was less expensive even when factoring in the cost of outsourcing the controlled packaging and associated shipping costs.
There were several challenges that impacted the results of our study. Because our team did not mandate the Rule of Five and rather allowed technicians to exercise some flexibility in their filling styles, we were unable to effectively measure the benefits of rounding during the active filling process.
The myriad forms of manufacturer’s packaging and ever-shifting drug availability also created challenges. Some manufacturer’s packs, such as ondansetron, furosemide, metoclopramide, came in cellophane-sealed trays while others were contained in blister packs of five (ie, ketorolac). In addition, standardizing the distribution process using the same manufacturer’s product consistently was complicated by the differing requirements from 340B and GPO contracts. Industry shortages resulting in back orders and formulation changes (ie, heparin) also caused drug availability challenges and variations in brand consistency.
The use of controlled packaging resulted in a statistically significant reduction in medication errors. Given the large number of medication doses dispensed in a busy hospital pharmacy, we have found that it is not uncommon for doses to be dispensed without a pharmacist first carefully checking for correct drug, dose, form, and expiration date. During the analysis of our controlled packaging approach, the pharmacist reviewer checked every dose for each of these parameters 100% of the time. While correct dose errors and expiration date errors occurred in our traditional dispensing approach, they were eliminated with the implementation of the controlled packaging approach. Given previous studies that have verified the direct correlation between the number of dispensing errors and adverse drug events (ADEs)1, we are confident that our new dispensing approach also has a positive impact on ADEs.
Designing Controlled Packaging
Poorly designed packaging can certainly impact safe distribution practices. We were particularly concerned with font sizes, adequate space placement, and color-coding guidelines and limitations. We found the most common font size used on SVP vial labels to be between four and six points. To remedy this, we created a general label template that used 13- to 16-point font sizes for the name of the drug, concentration, dose, and expiration date. The package’s specific contents (ie, number of vials), and the repackager’s identifying information and lot number were also included in the design. Because of frequently changing manufacturer sources, rather than use the NDC as the bar code on each package, we used the medication identification (Med ID) number assigned by the pharmacy computer system.
The size of the controlled package was designed to be large enough to hold up to five 10 mL vials—the largest vial size of most high-alert SVP drugs we commonly use—and to allow sufficient room to include the carousel label in cases where a drug order called for exactly five vials. The plastic material used for the bags was thin enough so that packages could be easily stored in carousel bins, with a perforated edge to expedite opening while filling the ADCs. We provided the final controlled package design template to our repackaging vendor.
Steven J. Ciullo, BPharm, MS, MPS, received his BS and master’s degree from St. John’s University School of Pharmacy, and his master’s degree in professional studies in health care administration from Long Island University and completed an ASHP Residency at Mercy Hospital. He is currently the director of pharmacy services at Upstate University Hospital, Upstate Medical University in Syracuse, New York.
Robert Wagner, RPh, received his BS in pharmacy from the University at Buffalo, School of Pharmacy and completed an ASHP Residency at E.J. Meyer Memorial Hospital. After 26 years with Eli Lilly and Company, Robert joined SUNY Upstate Medical University Hospital as a research pharmacist.
Patricia DeMasso-Anderson, MEd, received a BS in math education from SUNY Plattsburgh and a master’s in education from University of Illinois. Patricia began her career at SUNY as a research assistant and now serves as an investigational drug research assistant in the pharmacy department at Upstate University Hospital.
Acknowledgements and Disclosures:
Mr. Ciullo has registered the mark associated with Safe-T-Pack packaging with the U.S. Patent and Trademark Office but has no financial interest in any business associated with this trademark.
The authors would like to acknowledge the following Syracuse University, Martin J. Whitman School of Management faculty and students for their assistance: Gary LaPoint, assistant professor, Sarang Deshpande, MBA Class of 2011, Abhishek Shingwekar, MBA Class of 2011.
Funding for this project was provided by a grant from Health Research, Inc. (Contract Number 3677-01) on behalf of the New York State Department of Health and the New York State Attorney General’s Office as part of a settlement with Cardinal Health, Inc.